sign in sign up
<<
Home , Bioterrorism

PERSPECTIVE

Syndromic Surveillance and Bioterrorism-related Epidemics James W. Buehler,* Ruth L. Berkelman,* David M. Hartley,† and Clarence J. Peters‡

To facilitate rapid detection of a future bioterrorist Establishing a diagnosis is critical to the attack, an increasing number of public health departments response to a bioterrorism-related epidemic, since the are investing in new surveillance systems that target the diagnosis will guide the use of vaccinations, medications, early manifestations of bioterrorism-related disease. and other interventions. Absent a bioterrorism attack, pre- Whether this approach is likely to detect an epidemic soon- dicting whether syndromic surveillance will trigger an er than reporting by alert clinicians remains unknown. The detection of a bioterrorism-related epidemic will depend on investigation that yields a diagnosis before clinicians make population characteristics, availability and use of health and report a diagnosis is not possible. Our objective is to services, the nature of an attack, epidemiologic features of consider the mix of hypothetical factors that may affect the individual diseases, surveillance methods, and the capaci- detection of epidemics attributable to CDC category A ty of health departments to respond to alerts. Predicting bioterrorism agents (11). how these factors will combine in a bioterrorism attack may be impossible. Nevertheless, understanding their likely Establishing a Diagnosis effect on epidemic detection should help define the useful- Two pathways to establishing a diagnosis are described ness of syndromic surveillance and identify approaches to by the scenarios below and in Figure 1, using a single, clan- increasing the likelihood that clinicians recognize and report an epidemic. destine dissemination of an aerosol as an example.

Detection through Syndromic Surveillance ecause of heightened concerns about the possibility of The early signs of inhalational anthrax include nonspe- Bbioterrorist attacks, public health agencies are testing cific symptoms that may persist for several days before the new methods of surveillance intended to detect the early onset of more severe disease (12). Patients with prodromal manifestations of illness that may occur during a bioterror- illnesses seek outpatient care and are assigned nonspecific ism-related epidemic. Broadly labeled “syndromic surveil- diagnoses such as “viral syndrome.” Data on patients fit- lance,” these efforts encompass a spectrum of activities ting various syndromic criteria are transferred to the health that include monitoring illness syndromes or events, such department and tested for aberrant trends. This process as medication purchases, that reflect the prodromes of “flags” that a statistical detection threshold has been bioterrorism-related diseases (1–9). The Centers for exceeded. Epidemiologists conclude that a preliminary Disease Control and Prevention (CDC) estimates that, as investigation is warranted and collect blood for culture of May 2003, health departments in the United States have from several patients. Within 18 hours, one culture yields initiated syndromic surveillance systems in approximately a presumptive diagnosis of anthrax, prompting a full-scale 100 sites throughout the country (T. Treadwell, CDC, pers. response. comm.). The goal of these systems is to enable earlier detection of epidemics and a more timely public health Detection through Clinician Reporting response, hours or days before disease clusters are recog- Some persons in whom inhalational anthrax develops nized clinically, or before specific diagnoses are made and will have short incubation periods and prodromes (12). reported to public health authorities. Whether this goal is Respiratory distress occurs in one such person, and he is achievable remains unproved (4,5,10). hospitalized. Routine admission procedures include blood cultures. Within 18 hours, a presumptive diagnosis of anthrax is made. The patient’s physician informs the local *Emory University Rollins School of Public Health, Atlanta, health department, prompting a full-scale response. Georgia, USA; †University of Maryland School of Medicine, Baltimore, Maryland, USA; and ‡University of Texas Medical In practice, how a bioterrorism attack might be detect- Branch, Galveston, Texas, USA ed and diagnosed will probably be more complex.

Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 1197 PERSPECTIVE

firmed as positive for Bacillus anthracis. Two other A patients were seen by primary care physicians and sent home with diagnoses of viral syndrome or bronchitis 2–3 days before admission, including one patient who was begun on empiric therapy. For seven other patients, initial emergency room or hospital visits led directly to admission. In addition to the patient whose blood culture was obtained in an emergency room, seven others had not received prior antibiotic therapy, and B. anthracis was presumptively identified from blood within 24 hours of culture. One of these seven patients was the index patient, in whom B. anthracis was also recognized in cerebrospinal fluid within 7 hours of specimen collection. Three other patients had received before blood cultures were taken (one as an outpatient and two at the B time of hospital admission), requiring alternative diagnos- tic methods. Despite the small number of patients, their experience offers four lessons for detecting an epidemic of inhalation- al anthrax. First, a key objective of syndromic surveillance is to detect early-stage disease, but fewer than half of these patients sought care before hospitalization was necessary, and the interval between such care and admission was rel- atively narrow (1–3 days). This finding suggests that syn- dromic surveillance data must be processed, analyzed, and acted upon quickly if such data are to provide a clue to diagnosis in advance of late-stage disease. Second, emer- gency room data are a common source for syndromic sur- Figure 1. Number of cases of syndromic illness by time in a hypo- veillance, but detecting an increase in visits coincident thetical bioterrorism attack and two pathways to establishing a with hospital admission may not provide an early warning diagnosis: syndromic surveillance coupled with public health because the time needed to process surveillance data and investigation (upper pathway) and clinical and diagnostic evalua- investigate suspected cases would be at least as long as the tion of patients with short-incubation period disease (lower path- time for admission blood cultures to be positive for B. way). A, scenario favoring earlier detection by means of clinical evaluation. B, scenario favoring earlier detection by means of syn- anthracis. Blood cultures are likely to be routine for dromic surveillance. patients admitted with fever and severe respiratory illness, regardless of whether anthrax is considered as a diagnostic Published descriptions of 11 persons with inhalational possibility, and B. anthracis grows readily in culture in the anthrax in the United States in 2001 (13–19) provide some absence of prior antibiotic therapy, as observed in most of insight into this issue (Table 1 and Figure 2),1 even though these patients. Thus, if emergency room data are to be use- that epidemic was too small and geographically diffuse to ful in early detection of an anthrax epidemic, those data be detectable by syndromic surveillance. For six patients would need to be for visits that occur before hospital care with known dates of exposure, the median duration is required—a pattern observed in only two patients. Third, between exposure and symptom onset was 4 days (range the four patients who received early care and were dis- 4–6 days). The median duration between onset and the ini- charged to their homes were assigned three different diag- tial healthcare visit was 3 days (20) (range 1–7 days), and noses, which suggests that syndromic surveillance systems the median duration between onset of symptoms and hos- must address the potential variability in how patients with pitalization was 4 days (range 3–7 days). Two of the 11 the same may be diagnosed during the prodrome patients visited emergency departments and were sent phase. Fourth, rapid diagnosis after hospitalization was home with diagnoses of gastroenteritis or viral syndrome 1 possible only in those patients who had not received antibi- day before admission. In one patient, a blood culture obtained in the emergency room was read as positive for 1For interval calculations, if reported event dates were discrepant gram-positive bacilli the following day, which prompted in different case reports, dates reported by Jernigan et al. (13) recall of the patient. The culture was subsequently con- were used.

1198 Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 PERSPECTIVE

Table 1. Outcome of initial contact with health care for anthrax- tistical detection thresholds, availability of resources for related illness and timing of anthrax diagnosis, 11 patients with a followup, recent experience with false alarms, and criteria inhalational anthrax, 2001 for initiating investigations. Outcome No. of patients Disposition after initial medical care Season Admitted to hospital 7 A fifth key attribute is seasonality. An increase in ill- Discharged home from ER, subsequent hospital 2 admission ness associated with a bioterrorism attack may be more Discharged home from outpatient provider, 2 difficult to detect if it occurs during a seasonal upswing in subsequent hospital admission naturally occurring disease. Total 11 Agent- and disease-specific attributes may be among Anthrax diagnosis the most important factors affecting detection and diagno- Blood or CSF culture on hospital admission, 7 sis (Table 2). The incubation period and its distribution in presumptive diagnosis <24 h Blood culture from preceding ER visit, patient 1 the population will affect the rate at which new cases recalled for admission Prior antibiotic therapy; clinical suspicion of 3 anthrax; specialized test required to establish A diagnosis Total 11 aER, emergency room; CSF, cerebrospinal fluid. otics before cultures were taken. This finding emphasizes the importance of judicious use of antibiotics in patients with nonspecific illness. In addition to the specific attributes of individual Patient No. bioterrorism agents, multiple considerations will shape the recognition of a bioterrorism-related epidemic. Five of these attributes follow.

Size Syndromic surveillance would not detect outbreaks too Days after anthrax expossure small to trigger statistical alarms. Size would be affected by the virulence of the agent, its potential for person-to- B person transmission, the extent and mode of agent dissem- ination, whether dissemination occurs in more than one time or place, and population vulnerability.

Population Dispersion How persons change locations after an exposure will

affect whether disease occurs in a concentrated or wide Patient No. area, and thus whether clustering is apparent to clinicians or detectable through syndromic surveillance at specific sites.

Health Care Days after illness onset The more knowledgeable providers are about bioterror- ism agents, the greater the likelihood of recognition. Figure 2. Timeline to presumptive anthrax diagnosis, 11 patients Routine diagnostic practices or access to reference labora- with inhalational anthrax, 2001, United States. Abbreviations: Dx, tories may affect the timeliness of diagnosis for some dis- diagnosis; OutPt, outpatient visit followed by discharge home; ER, emergency room visit followed by discharge home. *Diagnosis eases. Familiarity with reporting procedures would delayed-initial blood cultures were negative in three patients who increase prompt reporting of suspected or diagnosed cases. received antibiotic therapy before culture specimens were collect- ed, requiring use of special diagnostic tests. For patients 1-10, Syndromic Surveillance case numbers correspond to those in report by Jernigan et al. (13); Syndromic surveillance will be affected by the selec- patient 11 reported by Barakat et al. (14). A, timeline begins with presumed date of anthrax exposure, available for six patients. B, tion of data sources, timeliness of information manage- timeline begins with day of illness onset for five patients without ment, definition of syndrome categories, selection of sta- recognized date of exposure

Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 1199 PERSPECTIVE

Table 2. Characteristics of bioterrorism-related epidemics that affect detection through clinical recognition versus syndromic surveillance Characteristicsa Clinical recognitionb Syndromic surveillancec Duration and variability of Broader distribution of incubation period More narrow distribution of incubation period—leading incubation period increases likelihood that patient with short to a steeper epidemic curve in the initial phase—increases incubation-period disease would be diagnosed likelihood that statistical threshold would be exceeded before a statistical threshold of syndromic cases sooner. is exceeded. Duration of nonspecific Shorter prodrome increases likelihood of Longer prodrome increases likelihood that increase in prodromal phase recognition or diagnosis at more severe or syndromic manifestations would be detectable and that fulminant stage. recognition of more severe stage (at which a diagnosis is more apt to be made) would be delayed. Presence or absence of clinical Presence increases likelihood of earlier clinical Absence decreases likelihood that diagnosis would be sign that would heighten recognition and diagnosis (e.g., mediastinal considered clinically, increasing opportunity for earlier suspicion of diagnosis widening on chest x-ray in inhalational anthrax). detection by means of syndromic surveillance. Likelihood of making diagnosis If diagnosis is apt to be made in the course of a If diagnosis is dependent on the use of a special test that in the course of routine routine diagnostic evaluation (not dependent on is unlikely to be ordered in the absence of clinical evaluation clinical suspicion of specific bioterrorism suspicion of diagnosis, then diagnosis in clinical care may infection), early diagnosis through clinical care is be delayed, increasing the opportunity for early detection likely. through syndromic surveillance. aInfection or disease attributes that may affect detection of an epidemic. bIncreases likelihood of initial detection through routine clinical care and reporting. cIncreases likelihood of initial detection through syndromic surveillance. develop (21) and thus how quickly an alarm threshold is tion period of a disease has a relatively narrow distribu- exceeded or whether clinicians recognize a temporal and tion, early recognition may offer little opportunity for post- geographic cluster. If a disease has a short prodrome, the exposure prophylaxis, although a potential benefit would chance is increased that a patient would be hospitalized remain for alerting healthcare providers and informing and a definitive evaluation initiated before an increase in their care of others with similar symptoms. This pattern of cases triggered a surveillance alarm. Alternatively, if a dis- illness is apt to result from exposure to an F. tularensis ease has a relatively long prodrome, chances are greater aerosol, which would likely result in an explosive epidem- that prediagnostic events (e.g., purchase of medications or ic with an abrupt onset and limited duration (22). use of outpatient care for nonspecific complaints) would accrue to levels that exceed syndromic surveillance thresh- Detecting Specific Bioterrorism olds, before definitive diagnostic evaluations are complet- Epidemics and Agents ed among patients with more severe disease. Arousing The attributes of the CDC category A bioterrorism clinical suspicion for a particular diagnosis will depend on agents that affect their detection, as well as the benefits of the specificity of both the early and late stages of illness as early detection, are summarized below, on the basis of well as the presence or absence of a typical feature that potential bioterrorism-related epidemic profiles devel- should alert clinicians to the diagnosis, such as mediastinal oped by experts (12,22,23,26–28). These profiles reflect widening in inhalational anthrax (12). If a routinely per- current knowledge of these diseases; their epidemiology formed test is apt to be diagnostic in a short time (e.g., the might differ if novel modes of dissemination or prepara- blood culture in anthrax), a rapid diagnosis is likely, even tion were employed. Each disease has attributes that could in the absence of clinical suspicion. If routine tests are increase or decrease the likelihood of early outbreak unlikely to yield a rapid diagnosis (e.g., the blood culture recognition through either clinical diagnosis or syndromic for the cause of , [22]), or surveillance. if the diagnosis requires a special test (e.g., the hemorrhag- ic fever [23]), a diagnosis may be delayed if not Inhalational Anthrax immediately considered. The distribution of the incubation period for inhalation- The public health benefit resulting from early detection al anthrax can be relatively broad as observed in of an epidemic is likely to vary by disease. If a disease has Sverdlovsk (2–43 days); most cases occur within 1–2 a relatively wide distribution of potential onsets, early weeks after exposure (24). In the 2001 U.S. outbreak, the recognition provides greater opportunity to administer pro- distribution of incubation periods was more limited, 4–6 phylaxis to exposed persons. For example, based on data days, although later-onset cases may have been averted by from the Sverdlovsk incident (24), Brookmeyer and antibiotic prophylaxis (25). The nonspecific prodrome for Blades estimated that use of antibiotic prophylaxis during anthrax may last from several hours to several days. Taken the 2001 anthrax outbreak prevented nine cases of inhala- together, these data suggest that the initial slope of an epi- tional disease among exposed persons (25). If the incuba- demic curve may be comparatively gradual during the first

1200 Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 PERSPECTIVE week, leading to slower recognition through syndromic diagnosis is suspected and special attention is given to surveillance than for other caused by bioterror- specimen processing. Confirming the diagnosis depends ist agents with pulmonary manifestations, such as on special tests available through reference laboratories. tularemia or (22,28). In contrast, medi- Treatment the first day of symptoms is generally consid- astinal widening on chest x-ray or computed tomographic ered necessary to prevent death in pneumonic plague, so scan or Gram stain of cerebrospinal or pleural fluid should early recognition of an aerosol plague attack would enable lead an alert and knowledgeable physician to consider the life-saving use of antibiotics in febrile patients and prophy- diagnosis of anthrax, even though these tests may not be laxis of contacts (28). conducted until relatively late in the clinical course. B. anthracis is likely to be detected quickly in cultures, favor- Botulism ing clinical recognition. Retrospective analysis of data Foodborne botulism typically has a relatively narrow from 2001 showed that inhalational anthrax can be distin- incubation period (12–72 hours), which may vary from 2 guished from influenzalike illness or community-acquired hours to 8 days, depending on the inoculum. For the three pneumonia by using an algorithm that combines clinical known cases of inhalational botulism attributed to a rela- and laboratory findings (20), although the practical utility tively low exposure to aerosolized , the incubation of this approach is untested. In addition to permitting period was approximately 72 hours (26). The characteris- antibiotic use among ill persons, early recognition would tic clinical picture of descending paralysis should prompt enable postexposure antibiotic prophylaxis (12,25). consideration of botulism, and this unique pattern among bioterrorism agents lends itself to a specific syndrome cat- Tularemia egory. However, the illness may be misdiagnosed, as The typical incubation period for tularemia is relatively observed in a large foodborne outbreak of botulism in narrow after a person is exposed to aerosolized F. tularen- 1985; 28 persons who had eaten at a particular restaurant sis, with abrupt onset of nonspecific febrile illness, with or and in whom botulism had developed were assigned other without respiratory symptoms, in 3–5 days (range 1–14 diagnoses before the geographically dispersed outbreak days), followed by rapid progression to life-threatening was recognized and publicized in the media (26,29). pneumonitis (22). This relatively narrow incubation period Symptoms of inhalational botulism, with choking, dyspha- for most patients and rapid progression to severe disease gia, and dysarthria dominating the clinical picture, may would lead to a rapid increase in cases after a large and differ from those associated with ingestion of toxin and acute exposure. Finding a number of such cases in a short complicate recognition of the disease. Specialized testing interval should trigger both syndromic surveillance alarms for is available at a limited number of and clinical suspicion. F. tularensis is a slow-growing and state laboratories and CDC. Postexposure prophylaxis is fastidious organism and may take up to 5 days after inoc- limited by the scarcity of, and potential for, allergic reac- ulation to be detectable, if it is detected at all, in a routine- tions to botulinum antitoxin, leading to recommendations ly processed blood culture. The use of special laboratory that exposed persons be observed carefully for early signs techniques may be required, delaying the likelihood of of botulism, which should prompt antitoxin use (26). detection in the absence of clinical suspicion. After an epi- Antitoxin should be given as early as possible, another fact demic is recognized, specific antibiotic therapy is recom- that highlights the importance of early detection. mended for exposed persons in whom a febrile illness Depending on the level of exposure and the geographic develops (22). dispersion of affected persons, syndromic surveillance for characteristic neurologic symptoms could aid outbreak Pneumonic Plague detection, or the occurrence of an epidemic might be obvi- Exposure to aerosolized results in pneu- ous to clinicians. monic plague, which has a typical incubation period of 2 to 4 days (range 1–6 days). The disease has a relatively short prodrome, followed by rapidly progressive pneumo- The incubation period of smallpox is usually 12–14 nia (28), which would lead to a rapid increase in cases at days but may range from 7 to 17 days. The early sympto- the onset of an epidemic. Standard clinical laboratory find- matic phase includes a severe febrile illness and appear- ings are nonspecific, which alone might not prompt clini- ance of a nonspecific macular rash over a 2- to 4-day peri- cal suspicion, but microscopic examination of a sputum od, followed by evolution to a vesicular and then pustular smear may show characteristic findings, which should rash over the next 4 to 5 days (27). Thus, the initial phase prompt consideration of the diagnosis. Cultures of blood or of smallpox may lend itself to detection through surveil- sputum are apt to show growth within 24 to 48 hours, but lance of a febrile rash illness syndrome. Once smallpox is routine procedures may misidentify Y. pestis unless the suspected, the can be rapidly detected by electron

Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 1201 PERSPECTIVE microscopic examination of vesicular or pustular fluid, if ly, such as pneumonic plague, syndromic detection of laboratory resources for electron microscopy are available, severe disease (e.g., through emergency room visits, hos- or by polymerase chain reaction, if the necessary primers pital admissions, or deaths) may be more feasible than are available. Contacts can be protected by vaccination up detection aimed at early indicators before care is sought to 4 days after exposure. Discourse is substantial about the (e.g., purchases of over-the-counter medications) or when relative merits of pre-event versus postevent vaccination illness is less severe (e.g., primary care visits). Whether (27,30–33). Syndromic surveillance may show an increase detection of syndromic late-stage disease offers an advan- in febrile rash illness, although once the characteristic rash tage over detection through clinical evaluation will depend appears, the diagnosis should be quickly established. on the attributes of the infections and diagnostic resources, as described above. Viral Hemorrhagic Fevers Predicting how the mix of relevant factors would com- This category includes multiple infectious agents that bine in a given situation to affect the recognition of a range from having a relatively broad to narrow incubation bioterrorism-related epidemic is difficult, although mathe- period (e.g., , 2–21 days; yellow fever 3–6 days). matical models may provide further insight (5). The most These diseases present with nonspecific prodromes that important factors affecting early detection are likely to be may have an insidious or abrupt onset. In severe cases, the the rate of accrual of new cases at the outset of an epidem- prodrome is followed by hypotension, shock, central nerv- ic, geographic clustering, the selection of syndromic sur- ous system dysfunction, and a bleeding diathesis. The dif- veillance methods, and the likelihood of making a diagno- ferential diagnosis includes a variety of viral and bacterial sis quickly in clinical practice. diseases. Establishing the diagnosis depends on clinical Ongoing efforts to strengthen the public health infra- suspicion and the results of specific tests that must be structure (34,35) and to educate healthcare providers about requested from CDC or the U. S. Army Medical Research bioterrorism diseases and reporting procedures should Institute of Infectious Diseases. The value of postexposure strengthen the ability to recognize bioterrorism outbreaks. prophylaxis with antiviral medications is uncertain, and For example, in New Jersey in 2001, reporting of two early (with the exception of yellow fever, for which a vaccine is cases of cutaneous anthrax was delayed until publicity available) response measures are limited to and about other anthrax cases prompted physicians to consider observation of exposed persons, treatment with ribavarin the diagnosis and notify the health department, suggesting (if the virus is one that responds to that antiviral drug), and that opportunities for earlier use of postexposure prophy- careful attention to infection control measures (23). laxis were missed (36). In addition, while the importance Patients seen with symptoms during the prodromal phase of new diagnostic tools, including rapid tests, should be may not clearly fit into a single syndrome category, but emphasized (37), the essential role of existing diagnostic syndromic surveillance focused on the early signs of a techniques should not be overlooked. Clinical suspicion is febrile bleeding disorder would be more specific. critical, and a key prompt for arousing clinical suspicion One of the biggest concerns about syndromic surveil- may be the microscopic examination of a routinely collect- lance is its potentially low specificity, resulting in use of ed specimen, as occurred in the index case of the 2001 resources to investigate false alarms (6,10). Specificity for anthrax outbreak, when a Gram stain of the cerebrospinal distinguishing bioterrorism-related epidemics from more fluid led to the diagnosis (15). However, as recently high- ordinary illness may be low because the early symptoms of lighted by the Institute of Medicine, the use of basic diag- bioterrorism-related illness overlap with those of many nostic tests has decreased because of efforts to reduce the common infections. Specificity for distinguishing any type costs of care, the increasing use of empiric broad-spectrum of outbreak from random variations in illness trends may antibiotic therapy, and federal laboratory regulations, such be low if statistical detection thresholds are reduced to as the Clinical Laboratory Improvement Amendments of enhance sensitivity and timeliness. The likelihood that a 1988, which have discouraged laboratory evaluation in given alarm represents a bioterrorism event will be low, some clinical settings (38). assuming that probability of such an event is low in a given While we have focused on the role of syndromic sur- locality. Approaches used to increase specificity include veillance in detecting a bioterrorism-related epidemic, requiring that aberrant trends be sustained for at least 2 other uses of syndromic surveillance include detecting nat- days or that aberrant trends be detected in multiple systems urally occurring epidemics, providing reassurance that epi- (2). Another approach to enhancing specificity would be to demics are not occurring when threats or rumors arise, and focus surveillance on the severe phases of disease, since tracking bioterrorism-related epidemics regardless of the the category A bioterrorism infections are more likely than mode of detection (4,6,10). Syndromic surveillance is many common infections to progress to life-threatening intended to enhance, rather than replace, traditional illness. For those diseases that are likely to progress rapid- approaches to epidemic detection. Evaluation of syn-

1202 Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 PERSPECTIVE dromic surveillance to consider the spectrum of potential 5. Committee on Science and Technology for Countering , uses is essential. A certain level of false alarms, as the National Research Council of the National Academies. Making the nation safer, the role of science and technology in countering - result of either syndromic surveillance or calls from clini- ism. Washington: National Academies Press; 2002. p. 74–6. cians, will be necessary to ensure that opportunities for 6. Henning KJ. Syndromic surveillance. In: Smolinski MS, Hamburg detection are not missed. Efforts to enhance the predictive MA, Lederberg J, editors. Microbial threats to health, emergence, value of syndromic surveillance will be offset by costs in detection, and response. Washington: Institute of Medicine National Academies Press; 2003. Appendix B. timeliness and sensitivity, and defining the right balance in 7. Lewis MD, Pavlin JA, Mansfield LJ, O’Brien S, Boomsma LG, practice, particularly in the absence of an accurate assess- Elbert Y, et al. Disease outbreak detection system using syndromic ment of bioterrorism risk, will be essential. data in the greater Washington DC area. Am J Prev Med Two committees of the National Academies have rec- 2002;23:180–6. 8. Broad WJ, Miller J. Health data monitored for bioterror warning. ommended more careful evaluation of the usefulness of New York Times, Jan 27, 2003, page A1. [Cited Feb 2003]. Available syndromic surveillance before it is more widely imple- from: URL: http://www.nytimes.com/2003/01/27/national/27DISE mented (5,38). Because the epidemiologic characteristics .html of different bioterrorism agents may vary in ways that 9. Centers for Disease Control and Prevention. Press release. CDC pro- vides $1.2 million to fund pilot program for early warning system for affect the detection of epidemics, these evaluations should terrorism-related illness outbreaks. [Cited Dec 2002] Available from: address the epidemiology of specific bioterrorism agents. URL: http://www.cdc.gov/od/oc/media/pressrel/r021002.htm Efforts to detect bioterrorism epidemics at an early stage 10. Reingold A. If syndromic surveillance is the answer, what is the ques- should not only address the development of innovative new tion? and Bioterrorism: Strategy, Practice, and Science 2003;1:1–5. surveillance mechanisms but also strengthen resources for 11. Rotz LD, Khan AS, Lillibridge SR, Ostroff SM, Hughes JM. Public diagnosis and enhance relationships between clinicians and health assessment of potential biological terrorism agents. Emerg public health agencies—relationships that will ensure that Infect Dis 2002;8:225–30. clinicians notify public health authorities if they suspect or 12. Inglesby TV, O’Toole T, Henderson DA, Bartlett JG, Ascher MS, Eitzen E, et al. Anthrax as a biological weapon, 2002, updated recom- diagnose a possible bioterrorism-related disease. mendations for management. JAMA 2002;287:2236–52. 13. Jernigan JA, Stephens DS, Ashford DA, Omenaca C, Topiel MS, Acknowledgments Galbraith M, et al. Bioterrorism-related inhalational anthrax: the first We wish to acknowledge the contributions of the anony- 10 cases reported in the United States. Emerg Infect Dis 2001;7:933–44. mous reviewers. 14. Barakat LA, Quentzel HL, Jernigan JA. Fatal inhalational anthrax in a 94-year-old Connecticut woman. JAMA 2002;287:863–8. Drs. Buehler and Berkelman were supported in part by a 15. Bush LM, Abrams BH, Beall A, Johnson CC. Index case of fatal grant from the O. Wayne Rollins Foundation. inhalational anthrax due to bioterrorism in the United States. N Engl J Med 2001;345:1607–10. Dr. Buehler is a research professor in the Department of 16. Borio L, Frank D, Mani V, Chiriboga C, Pollman M, Ripple M, et al. Epidemiology and a member of the Center for Public Health Death due to bioterrorism-related inhalational anthrax, report of 2 Preparedness and Research of the Rollins School of Public patients. JAMA 2001;286:2554–9. Health, Emory University. He serves as a consultant epidemiolo- 17. Mayer TA, Bersoff-Matcha S, Murphy C, Earls J, Harper S, Pauze D, et al. Clinical presentation of inhalational anthrax following bioter- gist to the Georgia Division of Public Health. His current rorism exposure, report of 2 surviving patients. JAMA research interests center on improving surveillance for the early 2001;286:2549–53. detection of bioterrorism-related epidemics. 18. Dewan PK, Fry AM, Laserson K, Tierney BC, Quinn CP, Hayslett JA, et al. Inhalational anthrax outbreak among postal workers, Washington, D.C., 2001. Emerg Infect Dis 2002:8:1066–72. References 19. Jernigan DB, Raghunathan PL, Bell BP, Brechner R, Bresnitz EA, Butler JC, et al. Investigation of bioterrorism-related anthrax, United 1. Lazarus R, Kleinman KP, Dashevsky I, DeMaria A, Platt R. Using States, 2001: epidemiologic findings. Emerg Infect Dis automated medical records for rapid identification of illness syn- 2002;8:1019–28. dromes (syndromic surveillance): the example of lower respiratory 20. Kuehert MJ, Doyle TJ, Hill HA, Bridges CB, Jernigan JA, Dull PM, infection. BMC Public Health 2001;1:9. et al. Clinical features that discriminate inhalational anthrax from 2. Mostashari F. BT surveillance in New York City. Presentation at the other acute respiratory illnesses. Clin Infect Dis 2003;36:328–36. CDC International Conference on Emerging Infectious Diseases, 21. Sartwell PE. The distribution of the incubation periods of disease 2002, Atlanta. [Cited January 2003] Available from: URL: (historical paper). Am J Epidemiol 1995;141:386–94. ftp://ftp.cdc.gov/pub/infectious_diseases/iceid/2002/pdf/mostashari. 22. Dennis DT, Inglesby TV, Henderson DA, Bartlett JG, Ascher MS, pdf Eitzen E, et al. Tularemia as a biological weapon, medical and public 3. Wagner MM, Tsui F, Espino JU, Dato VM, Sittig DF, Caruana RA, et health management. JAMA 2001;285;2763–73. al. The emerging science of very early detection of disease outbreaks. 23. Borio L, Inglesby T, Peters CJ, Schmaljohn AL, Hughes JM, Jahrling Journal of Public Health Management Practice 2001;7:51–9. PB, et al. Hemorrhagic fevers viruses as biological weapons: medical 4. Centers for Disease Control and Prevention. Draft framework for and public health management. JAMA 2002;287:2391–405. evaluating syndromic surveillance systems for bioterrorism prepared- 24. Meselson M, Guillemin J, Hugh-Jones M, Langmuir A, Popova I, ness. [Cited December 2002] Available from: URL: http:// Shelokov A, et al. The Sverdlovsk anthrax outbreak of 1979. Science www.cdc.gov/epo/dphsi/phs/syndromic.htm 1994;266:1202–7.

Emerging Infectious Diseases • Vol. 9, No. 10, October 2003 1203 PERSPECTIVE

25. Brookmeyer R, Blades N. Prevention of inhalational anthrax in the 35. United States Department of Health and Human Services. HHS US outbreak. Science 2002;295:1861. announces bioterrorism aid for states, including special opportunity 26. Arnon SS, Schechter R, Inglesby TV, Henderson DA, Bartlett JG, for advance fund. Press release. Mar 20, 2003 [Cited March 2003]. Ascher MS, et al. Botulinum toxin as a biological weapon: medical Available from: URL: http://www.hhs.gov/news/press/2003pres/ and public health management. JAMA 2001;285:1059–70. 20030320.html 27. Henderson DA, Inglesby TV, Bartlett JG, Ascher MS, Eitzen E, 36. Bresnitz EA, DiFerdinando GT. Lessons from the anthrax attacks of Jahrling PB, et al. Smallpox as a biological weapon: medical and pub- 2001, the New Jersey experience. Clinics in Occupational and lic health management. JAMA 1999;281:2127–37. Environmental Medicine 2003;2:227–52. 28. Inglesby TV, Dennis DT, Henderson DA, Bartlett JG, Ascher MS, 37. National Institute of Allergy and Infectious Diseases Office of Eitzen E, et al. Plague as a biological weapon: medical and public Communications and Public Liaison. HHS accelerates bioterrorism health management. JAMA 2000;283:2281–90. research: new programs expedite ideas from concerned scientists. 29. St Louis ME, Peck SH, Bowering D, Morgan GB, Blatherwick J, Press release. Dec 6, 2001 [Cited March 2003]. Available from: URL: Banerjee S, et al. Botulism from chopped garlic: delayed recognition http://www.niaid.nih.gov/newsroom/releases/accelbio.htm of a major outbreak. Ann Intern Med 1988;108:363–8. 38. Smolinski MS, Hamburg MA, Lederberg JA, editors. Microbial 30. Meltzer MI, Damon I, LeDuc JW, Millar JD. Modeling potential threats to health, emergence, detection, and response. Washington: responses to smallpox as a bioterrorist weapon. Emerg Infect Dis Institute of Medicine National Academies Press; 2003. p. 183–94. 2001;7:959–69. 31. Gani R, Leach S. Transmission potential of smallpox in contempo- Address for correspondence: James Buehler, Rollins School of Public rary populations. Nature 2001;414:748–51. 32. Kaplan EH, Craft DL, Wein LM. Emergency response to a smallpox Health, Rm. 416, Emory University, 1518 Clifton Rd., NE, Atlanta, GA attack: the case for mass vaccination. PNAS 2002:99:10935–40. 30322, USA; fax: 404-712-8345; email: [email protected] 33. Halloran ME, Longini IM, Nizam A, Yang Y. Containing bioterrorist smallpox. Science 2002;298:1428–32. 34. United States Department of Health and Human Services. HHS Use of trade names is for identification only and does not imply announces $1.1 billion in funding to states for bioterrorism prepared- endorsement by the Public Health Service or by the U.S. ness. Press release. Jan 31, 2002 [Cited March 2003]. Available from: Department of Health and Human Services. URL: http://www.hhs.gov/news/press/2002pres/20020131b.html

1204 Emerging Infectious Diseases • Vol. 9, No. 10, October 2003